The AI arms race drives up financing costs: US tech companies’ financing premium hits the highest level since 2003

The financing demand for AI infrastructure construction continues to surge, pushing up the debt financing costs of US high-tech companies. As US tech companies compete to spend hundreds of billions of dollars building data centers and computing infrastructure, the bond market has begun to demand higher risk compensation.

Data compiled by Bloomberg shows that the average bond financing cost for US investment-grade tech company borrowers, relative to the overall US market, has reached its highest premium since 2003.

The bond spreads for US tech companies have risen to the highest level in three years, while spreads in the overall investment-grade bond market remain relatively stable, further widening the gap in financing costs between the tech sector and other issuers.

US tech giants still need vast amounts of capital to maintain AI infrastructure construction, but as the scale of bond issuance continues to increase, investors are beginning to demand higher yields to compensate for potential credit risks and supply pressures. Analysts point out that as financing supply continues to expand, US tech bond spreads may further widen.

AI financing costs remain high, companies widely attempt cross-market financing

Data centers, chips, and power infrastructure require continuous, large-scale investment. Google, Microsoft, Meta, Amazon, and Oracle are among the major tech companies raising funds for related investments via the bond market.

US tech companies are trying to diversify their massive financing needs across different currencies and markets. BNP Paribas anticipates that investment-grade corporate bond net supply in the eurozone will rise by 83% to €192 billion ($214 billion) by 2027, up from €105 billion this year.

However, an increase in financing channels does not mean the market’s absorption capacity can expand without limit. As issuance volumes rise, investors are demanding greater risk compensation for new bonds.

AI financiers are relatively insensitive to financing prices, since they require large sums to maintain a competitive edge. Nevertheless, as more and more companies simultaneously tap the bond market, investors’ capital and risk tolerance have natural limits, and tech companies may be forced to offer higher yields to complete their financing smoothly.

Oracle and other firms’ credit risks draw attention

The bond market’s repricing of AI financing risk has already shown up in spreads and credit default swap (CDS) markets for some large tech firms.

Bond yields for hyperscale cloud service providers have risen recently, with Oracle’s bond spreads drawing particular attention. JPMorgan analysts note that Oracle’s bond spreads are about 2.8 times the broad investment-grade bond index, a record high.

The cost of credit risk hedging for AI-related companies is on the rise. Alphabet's CDS cost has nearly doubled in the past year, while Oracle’s CDS quotes have also climbed significantly.

According to the Financial Times, CDS market activity related to AI is increasing, with the net notional amount of such CDS rising from less than $7 billion a year ago to nearly $30 billion. Nvidia's five-year CDS spread has climbed to 141.5 basis points, reflecting growing investor concern about debt risks in the tech sector.

Changes in the CDS market do not signal imminent default by these companies, but they do indicate that investors are stepping up risk hedging and remaining more vigilant regarding the debt-servicing ability and future financing needs of companies within the AI industry chain.

AI investment returns and debt burdens become market focus

Rising financing costs for tech companies are underpinned by uncertainty between the scale of AI infrastructure investment and future cash flow returns.

Large tech firms are still accelerating the construction of data centers and expanding computing power, but these investments require continued capital, and the realization cycle for returns, the profitability of AI operations, and the size of future capital expenditures remain uncertain.

As debt financing scales up, investors are no longer just focused on whether companies can continue to raise capital, but are increasingly concerned about whether future cash flows can cover the increased debt burden.